SSTI commentary: What is a fair share of R&D? A closer look at benchmarking

Would you expect a community of 100,000 people to have less than one-half as much R&D activity as a community with 250,000 residents? Such a simple question cannot be considered without more information. You may ask which two communities are being compared. Would your answer be different if you learned the smaller community was a college town with a research-intensive university as its core economic engine, while the second community was largely a distribution hub and didn’t have a similar R&D asset?*  Yet politicians, pundits, media and even policymakers often benchmark cities, regions and states on incomplete or irrelevant  information.

Consider the varying share of the population in communities across your own state that are engaged in research or development. Would you expect R&D funding to be evenly distributed among those communities?  How, then, do or should we benchmark state or regional R&D intensity?

Policymakers, program designers and practitioners should consider how one improves benchmarks as indicators of progress for any policy or program, particularly when addressing equity issues. While some metrics lend themselves well to standardization on a per capita basis, R&D intensity and R&D capacity are not among them.

Per capita standardization of R&D persists, nonetheless, and at some peril for better public policy options. For instance, Science, the weekly magazine of the prestigious American Association for the Advancement of Science, recently included an article entitled “Does your state get its fair share of federal research dollars?” discussing a bill introduced in Congress that would require the National Science Foundation (NSF) to use a per capita formula for determining state eligibility to participate in the Experimental Program to Stimulate Competitive Research (EPSCoR). NSF presently determines eligibility based on the share of NSF funding awarded within each state – the current threshold is 0.75 percent of NSF funding. Twenty-five states, the Commonwealth of Puerto Rico, Guam, and the U.S. Virgin Islands meet the criterion.

The per capita versus state share measurement is a debate baked into American government and politics. What matters is how either measure —or alternatives — best maps to the policy objectives of the program in question. While at least relevant to the policy goal of increasing NSF-funded R&D activity in states that presently rank lowest on that measure, the NSF eligibility requirement assumes, like the per capita definition proposed, that high quality R&D can happen anywhere in the country and should/could be distributed evenly, unrelated to performance capacity.

If the current and proposed measures are inadequate, the introduction of the legislation offers the opportunity to consider better measures of “fair share” of R&D funding, as the Science article’s title asks. The goal for EPSCoR, according to NSF’s website, is to enhance “research competitiveness of targeted jurisdictions by strengthening STEM capacity and capability.” Share of NSF’s R&D dollars on a per capita or state basis does not seem to be a direct measure of this goal.

To be most fair, it would seem, the denominator of any standardizing ratio (in other words, R&D funding per what?) should be related to the numerator as closely as possible. NSF awards the vast majority of its funding, for example, to researchers within institutions of higher education – and only in select fields of science and engineering at that. Comparing NSF funding received in a state by a standardizing agent of its total S&E faculty, staff and graduate students, then, is certainly closer to a reasonable measure of fairness than total population.

That new measure also has limitations, however. One issue is the measure assumes all S&E faculty, staff and graduate students are equally eligible for each NSF dollar of funding when, in reality, the distribution of funding across S&E fields is far from level in the NSF budget. Additionally, each institution/state will not have the same number of researchers in each field, meaning they won’t all have the same chances to receive funding even if NSF’s budget was parsed evenly across fields. Thus, the standardizing measure must be even further refined to reflect the weighting of funding opportunity to possible funding recipient – a complex formula most likely.

To be certain any state is receiving its fair share of R&D expenditures, it is only fair to weigh the right factors.